Sunday, August 14, 2011

Cultural panic and the economics of riots

Striking events such as the recent England riots provide opportunities for all sorts of attempts to “explain” the events (in this case riots) in terms of whatever “cultural panic” framings one finds congenial.

Fortunately, riot and civil unrest has been the subject of some quantitative social science, which allow us to develop analysis based on actual empirical evidence.

Economists Jacopo Ponticelli and Hans-Joachim Voth have produced a paper (pdf) on whether fiscal austerity (i.e. budget cuts) lead to social unrest, examining the European evidence from 1919 onwards. The short answer is they do find quite a strong connection between budget cuts and riots and other form of civil unrest. Tax increases have little connection to civil unrest, but budget cuts do, independent of other changes in GDP. They also find, unsurprisingly, that a propensity to civil unrest is associated with higher levels of public debt. Given other evidence that budget cuts have little association with loss of votes, their suggestion that expenditure is higher to buy social peace seems plausible.

There is, however, an interesting complication in that the effect disappears from the late 1980s onward: in other words, the period of the “Great Moderation” in economics and the collapse of the Soviet bloc in 1989-91. That they find that strongly democratic societies (where there are considerable restraints on the executive) show less propensity to riot and civil unrest give reasons why the latter might matter.

They also find that strong levels of media penetration are not associated with increased civil unrest. So, standard “cultural panic” explanations such as de-regulation/“neoliberalism”; high levels of welfare spending; mass media are all “contra-indicated”, as social analysts say. (A useful summary and short discussion of their findings is here.)

Also contra-indicated, it would appear, is the fiscal austerity explanation as the connection disappears from the late 1980s onwards. That, however, OECD countries generally have experienced the biggest collapse in nominal spending since the 1930s, a return to 1930s conditions might lead to a return to 1930s patterns, given that the authors find a strong connection between budget cuts and civil unrest across the rest of the period. Though fiscal austerity was clearly not the only (or even dominant) cause of civil unrest.

The Ponticelli and Voth paper seeks to answer the question: “what connection, if any, is there between fiscal austerity and civil unrest?” and deals only with European cases. A paper (pdf) covering a much wider range of cases, and testing standard economic explanations of riots, was produced in 1994 by economists Denise DiPasquale and Edward L. Glaesar (i.e. after the 1992 L.A. riots): Glaesar has a sensible opinion piece based on their findings here.

They find (completely unsurprisingly) that riots are urban phenomena associated with higher levels of urbanization. (In particular, bigger cities are at more risk.) They also find that ethnic diversity and high levels of unemployment are associated with riots while homeownership makes them less likely to occur but does not affect their intensity once they start.

Lacking a job or not owning a home obviously gives folk less to lose from rioting while concentrated populations are more likely to achieve the necessary "critical mass" of people who believe they can “get away with it”. The Ponticelli and Voth finding that budget cuts do have a connection to civil unrest, but tax increases do not, fits in with the DiPasquale and Glaesar findings concerning unemployment and homeownership. Basically, the sort of people not inclined to riot—employed homeowners—pay taxes while the sort of people much more inclined to riot—unemployed renters—receive government expenditure.

The ethnic diversity effect fits in with Robert Putnam’s findings that ethnic/cultural diversity lowers social trust. An effect likely to be particularly strong if ethnic group A is, in effect, policing ethnic group B (e.g. a predominantly white police force policing a black area): an incident of perceived police brutality is a classic riot-triggering event.

DiPasquale and Glaesar found that migration and poverty are not associated with rioting, though lower GDP per capita countries are more likely to have riots (presumably because there are more people with less to lose). Dictatorships are less likely to have riots: which suggests that the Ponticelli and Voth finding about fiscal austerity not leading to civil unrest in Europe since the late 1980s is less about the spread of democracy and more about general economic conditions. (There may also have been less funding and activist support for causing trouble in non-communist states.)

None of this “proves” why a particular set of riots happen. But they are highly suggestive. What they suggest is that the 2011 English riots are classic ethnic-diversity, unemployment and police failure riots: a combination which is easily enough on its own to explain why the riots occurred when and where they did.

No comments:

Post a Comment